JP2005110995A - Blood sugar level measuring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To measure the blood sugar level non-invasively based on the temperature measurement. <P>SOLUTION: Different regression functions are used for a healthy person and a diabetic in stabilizing the measurement data by correcting the non-invasively measured blood sugar level based on the temperature measurement with the degree of saturation of oxygen in blood and the blood flow. The room temperature (environment temperature), the body surface temperature, the temperature change in a block in contact with the body surface, the temperature of radiation from the body surface, and the absorbance of at least two wavelengths are measured values necessary in this model. The parameter on the blood flow, radiation heat transfer quantity, and convection heat transfer quantity can be obtained from the temperature, and the parameter on the degree of saturation of oxygen in hemoglobin can be obtained from the absorbance. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a non-invasive blood sugar level measuring method and apparatus for measuring glucose concentration in a living body without collecting blood.

  Hilson et al. Reported that face and sublingual temperature change after intravenous injection of glucose in diabetic patients (Non-patent Document 1). Scott et al. Discuss the problem of thermoregulation with diabetic patients (Non-Patent Document 2). Based on these findings, Cho et al. Have proposed a method and apparatus for determining blood glucose concentration by measuring temperature without blood collection (Patent Documents 1 and 2).

Various attempts have been made to calculate the glucose concentration without blood collection. For example, near-infrared light of three wavelengths is irradiated to the measurement site to detect the transmitted light intensity, and the living body temperature is detected to obtain a representative value of the second derivative of absorbance, and the living body from a predetermined reference temperature is obtained. A method has been proposed in which the representative value is corrected in response to a temperature shift and a blood glucose concentration corresponding to the corrected representative value is obtained (Patent Document 3). In addition, heating or cooling is performed while monitoring the living body temperature at the measurement site, and the dimming degree based on light irradiation is measured at the moment when the temperature changes, and the glucose concentration causing the temperature dependence of the dimming degree is measured. (Patent Document 4). In addition, an apparatus has been reported that takes an output ratio between reference light and transmitted light after irradiating a sample, and calculates a glucose concentration from a linear expression of the logarithm of the output ratio and the temperature of the living body (Patent Document 5). . In addition, a method for measuring contributions of skin parameters such as melanin and skin thickness and correcting non-invasive measurements such as glucose concentration has been reported (Patent Document 6).
Diabete & Metabolisme, “Facial and sublingual temperature changes following intravenous glucose injection in diabetics” by RM Hilson and TDR Hockaday, 1982, 8, 15-19 Can. J. Physiol. Pharmacol., “Diabetes mellitus and thermoregulation”, by AR Scott, T. Bennett, IA MacDonald, 1987, 65, 1365-1376 US Pat. No. 5,924,996 US Pat. No. 5,795,305 JP 2000-258343 A JP-A-10-33512 JP-A-10-108857 US Pat. No. 5,725,480

  Glucose in the blood (blood glucose) is used for the glucose oxidation reaction in the cell and produces energy necessary for maintaining the living body. Especially in the state of basal metabolism, most of the generated energy is thermal energy to maintain body temperature, so it is expected that there is some relationship between blood glucose concentration and body temperature By the way. However, as is apparent when considering fever due to illness, body temperature varies depending on factors other than blood glucose concentration. Conventionally, a method for obtaining blood glucose concentration by temperature measurement without blood collection has been proposed, but it has been difficult to say that it has sufficient accuracy.

  An object of the present invention is to provide a method and an apparatus for obtaining blood glucose concentration with high accuracy without blood collection based on temperature data of a subject.

Blood glucose is supplied to cells throughout the body by the vascular system, particularly capillaries. Although there are complex metabolic pathways in the human body, glucose oxidation is basically a reaction in which blood sugar and oxygen react to produce water, carbon dioxide and energy. Here, oxygen is oxygen supplied from blood to cells, and the amount of oxygen supply is determined by the hemoglobin concentration in blood, hemoglobin oxygen saturation, and blood flow. On the other hand, the heat generated in the body by glucose oxidation is taken away from the body in the form of convection, heat radiation, conduction, and the like. Moreover, it is said that healthy people have better glucose oxidation efficiency than diabetic patients. From this, it can be considered that, particularly in a hyperglycemic state, the amount of heat production is higher in healthy subjects than in diabetic patients even if the blood glucose concentration is the same. We considered that body temperature is determined by the amount of energy produced by glucose combustion in the body, that is, the balance between heat production and heat dissipation.
(1) The amount of heat production and the amount of heat dissipation are regarded as equivalent.
(2) The amount of heat production is a function of blood glucose concentration and oxygen supply.
(3) The oxygen supply amount is determined by the blood hemoglobin concentration, the blood hemoglobin oxygen saturation, and the blood flow in the capillaries.
(4) The amount of heat dissipation is mainly determined by thermal convection and thermal radiation.
(5) The relationship between blood glucose concentration and heat production is different between diabetics and healthy subjects.

  According to this model, the body surface is subjected to heat measurement, and at the same time, parameters relating to oxygen concentration in the blood and parameters relating to blood flow are measured, and it has been found that blood glucose levels can be determined with high accuracy using these measurement results. Was completed. As an example, the measurement for obtaining the parameter can be performed on a part of a human body, for example, a fingertip. Parameters related to convection and radiation can be determined by measuring the fingertips thermally. The parameters relating to the blood hemoglobin concentration and the blood hemoglobin oxygen saturation can be obtained by spectroscopically measuring hemoglobin in the blood and determining the ratio of hemoglobin not bound to oxygen bound to oxygen. Parameters relating to blood flow can be obtained by measuring the amount of heat transfer from the skin.

  The blood glucose level measuring apparatus according to the present invention includes a measurement unit that acquires a plurality of measurement values regarding a body surface and a measurement environment, a selection unit for selecting whether a healthy person or a diabetic patient, and a plurality of measurement values acquired by the measurement unit. And a calculation unit that calculates a blood glucose level using a selection result obtained by the selection means. The apparatus further includes a storage unit that stores a plurality of regression functions, and the calculation unit reads the regression function corresponding to the selection result from the storage unit and performs calculation. More specifically, the storage unit stores a plurality of regression functions and average values and standard deviations of a plurality of parameters corresponding to the respective regression functions, and the calculation unit stores a regression function corresponding to the selection result, a parameter average value, and The standard deviation is read from the storage unit and the blood glucose level is calculated.

  The blood glucose level measuring apparatus according to the present invention also includes an input means for inputting the type of a healthy person or a diabetic patient, a plurality of temperatures derived from the body surface, and convective heat transfer and radiation related to heat dissipation from the body surface. A calorific value measuring unit that obtains information used for calculating the heat transfer amount, an oxygen amount measuring unit that obtains information related to the amount of oxygen in blood, and a plurality of measured temperatures and parameters corresponding to the amounts of oxygen in blood and blood glucose levels are associated with each other. As for the function, a storage unit separately stored in a function for healthy persons and a function for diabetic patients, and a plurality of measurement values input from the calorific value measuring unit and the oxygen content measuring unit are converted into the parameters, respectively, and input means According to the type inputted by the calculation unit, the parameter is applied to a function for a healthy person or a function for a diabetic patient stored in the storage unit, and a blood glucose level calculated by the calculation unit And a display unit that displays. The storage unit stores a regression function for healthy persons and a regression function for diabetic patients, and the calculation unit calculates a blood glucose level using a regression function corresponding to the input type. More specifically, the storage unit stores a regression function for healthy persons, a regression function for diabetic patients, and an average value and standard deviation of a plurality of parameters included in each regression function. The blood glucose level is calculated using the regression function corresponding to the input type and the average value and standard deviation of the parameters associated with the regression function. The oxygen amount measurement unit includes a blood flow measurement unit that obtains information related to the blood flow rate, and an optical measurement unit that obtains hemoglobin concentration and hemoglobin oxygen saturation in the blood.

  The blood glucose level measurement method according to the present invention includes a step of acquiring a plurality of measurement values regarding a body surface and a measurement environment, a step of acquiring a type of a healthy person or a diabetic patient, a plurality of acquired measurement values and Calculating a blood glucose level using a regression function or a regression function for diabetic patients. The step of calculating the blood glucose level includes a step of obtaining a plurality of parameters from the obtained plurality of measured values, a step of normalizing the obtained plurality of parameters with an average value and a standard deviation corresponding to a healthy person or a diabetic patient, And applying a plurality of normalized parameters to a regression function corresponding to a normal person or a diabetic patient.

  ADVANTAGE OF THE INVENTION According to this invention, a highly accurate non-invasive blood glucose level measuring apparatus and measuring method can be provided.

Embodiments of the present invention will be described below with reference to the drawings.
First, the embodiment of the model will be described. Here, in consideration of (5) of the above-mentioned model, the embodiment corresponding to each of the diabetic patient and the healthy person is performed.

  Considering the amount of heat dissipation, convective heat transfer, the main factor, is related to the temperature difference between the ambient temperature (room temperature) and the body surface temperature, and the other main factor, the amount of heat dissipation due to radiation, is Stefan.・ It is proportional to the fourth power of body surface temperature according to Boltzmann's law. Therefore, it can be seen that the amount of heat dissipated from the human body is related to the room temperature and the body surface temperature. On the other hand, the amount of oxygen supply, which is another factor related to the amount of heat production, is expressed as the product of hemoglobin concentration, hemoglobin oxygen saturation, and blood flow.

  Here, the hemoglobin concentration can be measured from the absorbance at a wavelength (equal absorption wavelength) at which the molar extinction coefficients of oxygen-bound hemoglobin and reduced (deoxygenated) hemoglobin are equal. The hemoglobin oxygen saturation is determined by measuring the absorbance at the above-mentioned equiabsorption wavelength and the absorbance at one other wavelength with a known ratio of the molar extinction coefficient of oxygen-binding hemoglobin and reduced (deoxygenated) hemoglobin. Can be measured by solving That is, the hemoglobin concentration and the hemoglobin oxygen saturation can be obtained by measuring the absorbance of at least two wavelengths.

  What remains is the blood flow. The blood flow rate can be measured by various methods. An example of the measurement method will be described below.

FIG. 1 is a model diagram illustrating heat transfer from a body surface to a block when a solid block having a certain heat capacity is released from contact with the body surface for a certain period of time. The block material may be a resin such as plastic, for example, vinyl chloride. Here, attention is focused on the time change of the temperature T ₂ temporal change of the temperature T ₁ of the portion in contact with the body surface of the block, at a position away from the body surface on the block. The blood flow rate can be estimated mainly by tracking the time change of the temperature T ₂ (the temperature at a spatially separated point on the block). Details will be described below.

Before the block comes into contact with the body surface, the temperatures T ₁ and T ₂ at the two points of the block are equal to the room temperature _Tr . When the body surface temperature T _s is higher than the room temperature T _r , when the block comes into contact with the body surface, the temperature T ₁ quickly rises due to heat transfer from the skin and approaches the body surface temperature T _s . On the other hand, the temperature T ₂ is attenuated more than T ₁ and rises gently because the heat conducted in the block is radiated from the block surface. The temporal changes in the temperatures T ₁ and T ₂ depend on the amount of heat transfer from the body surface to the block. The amount of heat transfer from the body surface to the block depends on the blood flow in the capillaries flowing under the skin. Considering a capillary as a heat exchanger, the heat transfer coefficient from the capillary to the surrounding cellular tissue is given as a function of blood flow. Therefore, if the amount of heat transfer from the body surface to the block is measured by tracking the time changes of the temperatures T ₁ and T ₂ , the amount of heat transfer from the capillaries to the cellular tissue can be estimated, and the blood flow rate can be estimated from this. I can do it. Therefore, if the amount of movement from the body surface to the block is measured by temporally tracking the temperature change of T ₁ and T _2, the amount of heat transfer from the capillaries to the cellular tissue can be estimated, and the blood flow rate can be estimated from this. Can be estimated.

FIG. 2 is a diagram showing temporal changes in measured values of the temperature T _{1 of} the part in contact with the body surface in the block and the temperature T _{2 at} a position on the block away from the body surface contact position. When contacting the block to the body surface T ₁ measured value swiftly rises, and it gradually drops as release.

The Figure 3 illustrates a time variation of the measured value of the temperature T ₃ measured by a radiation temperature detector. Since the temperature T ₃ for measuring the temperature due to radiation from the body surface, it is more sensitive to temperature changes than other sensors. Since radiant heat propagates as electromagnetic waves, temperature changes can be transmitted instantaneously. Therefore, for example, as shown in FIG. 7 to be described later, is set near where the block contacts the body surface location to a radiation temperature detector for detecting the radiation heat from the body surface, the block and the body surface from the change in temperature T ₃ The contact start time t _start and the contact end time t _end can be detected. For example, as shown in FIG. 3, a temperature threshold value is set, and when the temperature threshold value is exceeded, the contact start time t _start is set, and when the temperature threshold value is lowered, the contact end time t _end is set. The temperature threshold is set to a temperature such as 32 ° C., for example.

Subsequently, the T ₁ measurement value between time t _start and time t _end is approximated by an S-shaped curve, for example, a logistic curve. The logistic curve is expressed by the following equation, where T is temperature and t is time.

The measured values can be approximated by obtaining the coefficients a, b, c, and d by the nonlinear least square method. A value obtained by integrating T with respect to the obtained approximate expression from time t _start to time t _end is defined as S ₁ .

Similarly, an integrated value S ₂ is calculated from the T ₂ measured value. At this time, the smaller (S ₁ -S ₂ ), the greater the amount of heat transfer from the body surface to the position of T ₂ . In addition, (S ₁ -S ₂ ) increases as the body surface contact time t _CONT (= t _end −t _start ) increases. Therefore, the a ₅ as a proportionality coefficient, a ₅ / a _{(t CONT × (S 1 -S} 2)) to indicate the parameter _{X 5} blood flow.

  From the above explanation, the measurement amount necessary for obtaining the blood glucose concentration by the model is room temperature (environmental temperature), body surface temperature, temperature change of the block in contact with the body surface, temperature due to radiation from the body surface. It can be seen that the absorbance is at least two wavelengths.

FIG. 4 is an explanatory diagram illustrating the relationship between measured values obtained by various sensors and parameters derived therefrom. A block in contact with the body surface is prepared, and two types of temperature T ₁ and T ₂ with time are measured by two temperature sensors installed at two locations. Separately, to measure the radiation temperature _{T 3} and the room temperature _{T 4} of the body surface. Further, the absorbances A ₁ and A ₂ are measured at at least two types of wavelengths related to the absorption of hemoglobin. Parameters relating to blood flow are obtained from the temperatures T ₁ , T ₂ , T ₃ , T ₄ . It provides a parameter related amount of heat transferred by radiation from the temperature T _3, provide parameters related to the amount of heat transferred by convection from the temperature T ₃ and the temperature T _4. Provides a parameter related to the hemoglobin concentration from the absorbance A _1, parameters are obtained from the absorbance A ₁ and A ₂ relates to the hemoglobin oxygen saturation.

  Next, a specific apparatus configuration for realizing non-invasive blood sugar level measurement according to the principle of the present invention will be described.

  FIG. 5 is a top view of the non-invasive blood sugar level measuring apparatus according to the present invention. In this device, the skin of the belly of the fingertip is used as the body surface, but other body surfaces can be used.

  On the upper surface of the apparatus, there are provided an operation section 11, a measurement section 12 on which a finger to be measured is placed, a display of measurement results, a display section 13 for displaying the state of the apparatus and measured values. The operation unit 11 has four push buttons 11a to 11d for operating the apparatus. The measurement unit 12 is provided with a cover 14, and when the cover 14 is opened (the figure shows a state in which the cover is opened), there is a finger placement unit 15 having an elliptical periphery. In the finger placement unit 15, there are an opening end 16 of the radiation temperature sensor unit, a contact temperature sensor unit 17, and an optical sensor unit 18.

  FIG. 6 shows a functional block diagram of the apparatus. This apparatus is driven by a battery 41. Signals measured by the sensor unit 48 composed of a temperature sensor and an optical sensor enter the analog / digital converters AD1 to AD5 installed corresponding to the respective signals and are converted into digital signals. The LED selection LSI 19 has a function of causing two light emitting diodes, which are light sources of the optical sensor, to emit light in a time-sharing manner under the control of the microprocessor 55. Peripheral circuits of the microprocessor 55 include analog / digital converters AD1 to AD5, a liquid crystal display 13, an LED selection LSI 19, a RAM 42, an IC card 43, and a real-time clock 45. These are connected to the micro via the bus lines 44. Accessed from the processor 55. The push buttons 11a to 11d are connected to the microprocessor 55, respectively.

  The microprocessor 55 has a built-in ROM for storing software. The microprocessor 55 has an interrupt factor register and an interrupt mask register as registers relating to interrupt processing requests input by pressing the buttons 11a to 11d. The interrupt factor register is a register for identifying a button pressed when an interrupt processing request is made to the microprocessor 55. The interrupt mask register is composed of 1 bit. By setting this register to 1 by software, acceptance of an interrupt processing request from a push button input can be prevented, that is, masking can be performed. When the software clears this register to 0, the interrupt processing request mask is canceled. As described above, this software performs control of various registers, access to ROM storage information and selection of ROM storage information according to a request input by a button, and computation using the ROM storage information. Further, as will be described later, the present microprocessor includes functions of a selection unit and a calculation unit.

  7A and 7B are diagrams showing a detailed example of the measurement unit, in which FIG. 7A is a top view, FIG. 7B is an XX sectional view thereof, and FIG. 7C is a YY sectional view thereof.

  First, temperature measurement by the non-invasive blood sugar level measuring apparatus of the present invention will be described. A thin plate 21 made of a material having high thermal conductivity, for example, gold, is disposed at a portion where the test portion (finger's belly) is in contact, and the thermal conductivity is higher than that of the plate 21 thermally connected to the plate 21. A rod-like heat conducting member 22 made of a low material, such as polyvinyl chloride, extends inside the apparatus. As the temperature sensor, the temperature of the plate 21 is measured, and the temperature of the thermistor 23 that is a temperature detector adjacent to the portion to be tested and the portion of the heat conducting member that is a fixed distance away from the plate 21 are measured, A thermistor 24 that is an indirect temperature detector is provided for the portion to be examined. An infrared lens 25 is disposed at a position inside the apparatus through which the subject portion (finger's belly) placed on the finger placement portion 15 can be seen, and a thermopile 27 is disposed below the infrared lens 25 via an infrared transmission window 26. . Further, another thermistor 28 is installed in the vicinity of the thermopile 27.

As described above, the temperature sensor unit of the measurement unit has four temperature sensors, and measures the following four types of temperatures.
(1) Finger surface temperature (thermistor 23): T ₁
(2) Temperature of the heat conducting member (thermistor 24): T ₂
(3) Radiation temperature of finger (thermopile 27): T ₃
(4) Room temperature (thermistor 28): T ₄

  Next, the optical sensor unit 18 will be described. The optical sensor unit is for measuring the hemoglobin concentration and the hemoglobin oxygen saturation necessary for obtaining the oxygen supply amount. In order to measure the hemoglobin concentration and hemoglobin oxygen saturation, it is necessary to measure absorbance at a minimum of two wavelengths, and FIG. 7C performs two-wavelength measurement using two light sources 33 and 34 and one detector 35. An example of the configuration is shown.

  The ends of the two optical fibers 31 and 32 are located in the optical sensor unit 18. The optical fiber 31 is an optical fiber for irradiating light, and the optical fiber 32 is an optical fiber for receiving light. As shown in FIG. 7C, the optical fiber 31 is connected to fibers 31a and 31b serving as branch lines, and light emitting diodes 33 and 34 having two wavelengths are disposed at the ends thereof. A photodiode 35 is disposed at the end of the light receiving optical fiber 32. The light emitting diode 33 emits light having a wavelength of 810 nm, and the light emitting diode 34 emits light having a wavelength of 950 nm. The wavelength 810 nm is an equiabsorption wavelength at which the molar extinction coefficients of oxygen-bound hemoglobin and reduced (deoxygenated) hemoglobin are equal, and the wavelength 950 nm is a wavelength at which the difference between the molar extinction coefficients of oxygen-bound hemoglobin and reduced hemoglobin is large. It is.

The two light-emitting diodes 33 and 34 emit light in a time-sharing manner, and light generated from the light-emitting diodes 33 and 34 is applied to the subject's finger from the optical fiber 31 for light irradiation. The light irradiated to the finger is reflected by the skin of the finger, enters the light receiving optical fiber 32, and is detected by the photodiode 35. When light applied to the finger is reflected by the skin of the finger, part of the light enters the tissue through the skin and is absorbed by hemoglobin in the blood flowing through the capillaries. The data measured by the photodiode 35 is the reflectance R, and the absorbance is approximately calculated by log (1 / R). Irradiation is performed for light having a wavelength of 810 nm and a wavelength of 950 nm, R is measured for each, and log (1 / R) is obtained, whereby absorbance A ₁ at a wavelength of 810 nm and absorbance A _{2 at a} wavelength of 950 nm are measured.

Assuming that the reduced hemoglobin concentration is [Hb] and the oxygen-binding hemoglobin concentration is [HbO ₂ ], the absorbance A ₁ and the absorbance A ₂ are expressed by the following equations.

A _Hb (810 nm), A _Hb (950 nm), A _{HbO 2} (810 nm) and A _{HbO 2} (950 nm) are molar extinction coefficients of reduced hemoglobin and oxygen-bonded hemoglobin, respectively, and are known at each wavelength. a is a proportionality coefficient. The hemoglobin concentration [Hb] + [HbO ₂ ] and the hemoglobin oxygen saturation [HbO ₂ ] / ([Hb] + [HbO ₂ ]) are obtained from the above equation as follows.

  Although an example of measuring hemoglobin concentration and hemoglobin oxygen saturation by measuring absorbance at two wavelengths has been described here, measuring the absorbance at three wavelengths or more can reduce the influence of interference components and increase the measurement accuracy. Is possible.

  FIG. 8 is a conceptual diagram showing the flow of data processing in the apparatus. In the apparatus of this example, there are five sensors including a thermistor 23, a thermistor 24, a thermopile 27, a thermistor 28, and a photodiode 35. Since the photodiode 35 measures the absorbance at a wavelength of 810 nm and the absorbance at a wavelength of 950 nm, six types of measurement values are input to the apparatus.

The five types of analog signals are converted into digital signals by analog to digital converters AD1 to AD5 via amplifiers A1 to A5, respectively. Parameters x _{i (i} = 1,2,3,4,5) is calculated from the digitally converted values. Specifically, x _i is expressed as follows. (A ₁ ~a ₅ is a proportionality coefficient)

Then, it calculates a normalized parameter from the mean and the standard deviation of the parameter x _i. A normalization parameter X _i (i = 1, 2, 3, 4, 5) is calculated from each parameter x _i by the following equation.

  With the above-mentioned five normalization parameters, conversion calculation to glucose concentration for final display is performed. A program necessary for processing calculation is stored in a ROM built in a microprocessor incorporated in the apparatus. Similarly, a memory area necessary for processing calculation is secured in a RAM incorporated in the apparatus. The calculation result is displayed on the liquid crystal display.

  The ROM contains a regression function for obtaining the glucose concentration C as a program component necessary for processing calculation. The regression function is obtained by the least square method using the glucose concentration measured by the enzyme electrode method, which is an invasive method for a large number of diabetic patients and healthy persons, and the normalization parameter obtained simultaneously for the large number of diabetic patients and healthy persons. Are determined in advance for each healthy person. When a common regression function determined from a measurement data group in which diabetic patients and healthy subjects are mixed, the relationship between the blood glucose concentration and the amount of heat production in the above model is as follows. Considering the point, it is considered that the correlation coefficient with the glucose concentration by the enzyme electrode method becomes small. Therefore, a separate regression function for each diabetic patient and healthy person is determined from each of the diabetic patient data group and the healthy person data group, and stored in the ROM.

Hereinafter, a method for determining a regression function will be described using a regression function for diabetic patients as an example. First, the glucose concentration C is expressed by the following formula (1). a _Di (i = 0,1,2,3,4,5) is determined in advance from measurement data of a large number of diabetic patients. The procedure for _obtaining a _Di is as follows.
(1) A multiple regression equation indicating the relationship between the normalization parameter and the glucose concentration C is created.
(2) A normal equation (simultaneous equation) relating to a normalization parameter is obtained from an equation obtained by the method of least squares.
(3) The value of the coefficient a _Di (i = 0, 1, 2, 3, 4, 5) is obtained from the normal equation and substituted into the multiple regression equation.

First, the following regression equation (1) showing the relationship between the glucose concentration C and the normalization parameters X _D1 , X _D2 , X _D3 , X _D4 , and X _D5 of the diabetic patient is created.

Subsequently, the least square method is used to obtain a multiple regression equation that minimizes an error from the glucose concentration measurement value C _Di by the enzyme electrode method. If the residual sum of squares is _RD , _RD is expressed by the following equation (2).

残差の二乗和Ｒ_Dが最小になるのは、式（２）をａ_D0，ａ_D1，…，ａ_D5で偏微分してゼロとなるときなので、次式が得られる。

The residual sum of squares R _D is minimized when the equation (2) is partially differentiated by a _D0 , a _D1 ,..., A _D5 and becomes zero.

C_D、Ｘ_D1〜Ｘ_D5の平均値をＣ_Dmean、Ｘ_D1mean〜Ｘ_D5meanとするとＸ_Dimean＝０（ｉ＝１〜５）であるので、式（１）から式（４）が得られる。

If the average values of C _D and X _{D1 to} X _D5 are C _Dmean and X _{D1mean to} X _D5mean , X _Dimean = 0 (i = 1 to 5), and therefore, Expression (4) is obtained from Expression (1).

また、正規化パラメータ間の変動・共変動は、式（５）で表され、正規化パラメータＸ_Di（ｉ＝１〜５）とＣとの共変動は式（６）で表される。

Further, the fluctuation / covariation between the normalization parameters is expressed by Expression (5), and the covariation between the normalization parameter X _Di (i = 1 to 5) and C is expressed by Expression (6).

式（４）（５）（６）を式（３）に代入して整理すると、連立方程式（正規方程式）（７）が得られ、これを解くことでａ_D1〜ａ_D5が求まる。

Substituting equations (4), (5), and (6) into equation (3) and rearranging results in simultaneous equations (normal equations) (7), which are solved to obtain a _{D1 to} a _D5 .

定数項ａ_D0は、式（４）を用いて求める。以上で求めたａ_Di（i=0,1,2,3,4,5）は装置製造時にＲＯＭに格納される。装置による実際の測定では、測定値から求めた正規化パラメータＸ_D1〜Ｘ_D5を回帰式（１）に代入することで、グルコース濃度Ｃが算出される。

The constant term a _D0 is obtained using equation (4). The a _Di (i = 0, 1, 2, 3, 4, 5) obtained above is stored in the ROM when the apparatus is manufactured. In actual measurement by the apparatus, the glucose concentration C is calculated by substituting the normalization parameters X _{D1 to} X _D5 obtained from the measurement values into the regression equation (1).

Similarly, the coefficient a _Ni (i = 0,1,2,3,4,5) for the healthy person is determined in advance from the measurement data of a large number of diabetic patients, and the ROM is obtained as a regression function (8) for the healthy person. Stored in

以下にグルコース濃度の算出過程の具体例を示す。予め糖尿病患者に対して測定した多数のデータから回帰式（１）の係数が決められており、マイクロプロセッサのＲＯＭには下記に示すグルコース濃度の算出式（９）が格納されている。さらに、パラメータｘ_１〜ｘ_５の平均値と標準偏差が格納されている。

A specific example of the glucose concentration calculation process is shown below. The coefficient of the regression equation (1) is determined from a large number of data measured in advance for diabetic patients, and the calculation formula (9) for the glucose concentration shown below is stored in the ROM of the microprocessor. Further, an average value and a standard deviation of the parameters x _{1 to} x ₅ are stored.

同様に、健常者に対するグルコース濃度の算出式（１０）及びパラメータｘ_１〜ｘ_５の平均値と標準偏差がＲＯＭに格納されている。

Similarly, the calculation formula (10) of the glucose concentration for the healthy person and the average value and standard deviation of the parameters x _{1 to} x ₅ are stored in the ROM.

Ｘ_D1〜Ｘ_D5はパラメータｘ_１〜ｘ_５を糖尿病患者の平均値及び標準偏差で正規化したものである。Ｘ_N1〜Ｘ_N5はパラメータｘ_１〜ｘ_５を健常者の平均値及び標準偏差で正規化したものである。パラメータの分布が正規分布であると仮定すると、正規化パラメータの95％は−２から＋２の間の値をとる。

X _{D1 to} X _D5 are parameters x _{1 to} x ₅ normalized by the mean value and standard deviation of diabetic patients. X _{N1 to} X _N5 are values obtained by normalizing the parameters x _{1 to} x ₅ with the average value and standard deviation of healthy subjects. Assuming that the parameter distribution is a normal distribution, 95% of the normalized parameters take values between -2 and +2.

As an example of measured values for diabetic patients, the normalization parameters X ₁ = + 0.15, X ₂ = -0.10, X ₃ = -0.22, X ₄ = -0.11, X ₅ = -0.09 are expressed in the above equation (9). Substitution results in C = 220 mg / dl. As an example of the measurement value of a healthy person, the normalization parameters X _N1 = −0.05, X _N2 = + 0.03, X _N3 = + 0.06, X _N4 = −0.10, X _N5 = + 0.12 are expressed by the above formula (10 Substituting into () results in C = 94 mg / dl.

Hereinafter, the details of the measurement using a device in which the calculation formula of the glucose concentration and the average values and standard deviations of the parameters x _{1 to} x ₅ of the diabetic patient and the healthy person are stored in the ROM will be described.

  FIG. 9 shows the operation procedure of the apparatus. The operation procedure of the apparatus includes a step of selecting a diabetic patient and a healthy person by a selection unit. When the button of the operation unit which is the selection input means is pressed to turn on the power of the device, “warming up” is displayed on the liquid crystal display, and the electronic circuit in the device is warmed up. At the same time, a check program runs and automatically checks the electronic circuit. When the “warming up” is completed, “Is you diabetic?” Is displayed on the liquid crystal display, and the subject is requested to select whether or not the subject is diabetic. The push button 11d corresponds to “Yes”, and the push button 11a corresponds to “No”. In accordance with the contents input here, the software stored in the ROM selects a function for obtaining an average value / standard deviation and a glucose concentration necessary for calculating a normalization parameter described later. The detailed operation flow of this selection unit will be described later. After the “Yes” or “No” button is pressed by the subject, “Please put your finger” is displayed on the liquid crystal display. When a finger is placed on the finger placement unit, a countdown is displayed on the liquid crystal display. When the countdown is finished, “Please release your finger” is displayed on the LCD. When the finger is removed from the finger placement unit, “data processing in progress” is displayed on the liquid crystal display. Thereafter, the blood glucose level is displayed on the liquid crystal display. At this time, the displayed blood glucose level is stored in the IC card 43 together with the date / time. After reading the displayed blood glucose level, press the button on the operation unit. After about 1 minute, the device waits for the next measurement and “Place your finger” is displayed on the LCD.

  FIG. 10 shows a detailed operation flow of the selection unit described above. As shown in FIG. 10, the selection unit indicates a mechanism for selecting either a diabetic patient or a healthy person. After warming up, an input request message indicating whether the subject is a diabetic is displayed on the liquid crystal display. The input request message requests the input by assigning the push button 11d to “Yes” and 11a to “No”. Thereafter, the software clears the interrupt mask register to 0, and enables input of an interrupt processing request by a button pressing operation. Subsequently, the microprocessor itself is shifted to a standby state in order to suppress current consumption. When any button is pressed by the subject, an interrupt processing request is issued to the microprocessor, and the microprocessor returns from the standby state. After returning from the standby state, the software sets the interrupt mask register to 1, and masks the interrupt processing request due to the button pressing operation. Subsequently, the software reads the interrupt factor register and determines which button has been pressed. When the button 11d is pressed, since the subject is a diabetic patient, the software loads the regression function for the diabetic patient and the average value / standard deviation of each parameter for the diabetic patient from the ROM. When the button 11a is pressed, since the subject is a healthy person, the software loads the regression function for the healthy person and the average value / standard deviation of each parameter for the healthy person from the ROM. The button 11b or 11c is for inputting another arbitrary measurement condition. When these buttons are pressed, the button 11b or 11c is invalid and returns to the input request to the subject. The detailed operation flow of the selection unit has been described above, and blood glucose measurement is started after the selection unit finishes selecting either a diabetic patient or a healthy person.

The measurement results obtained by the enzyme electrode method in which blood obtained by blood collection, which is a conventional measurement method, is reacted with a reagent and the amount of electrons generated by this reaction is measured to measure the blood glucose level, and the measurement result according to one embodiment of the present invention Is described below. As an example of the measured value of a diabetic patient, when the glucose concentration by the enzyme electrode method is 236 mg / dl, the normalization parameters X ₁ = + 0.15, X ₂ = −0.10, X ₃ = -0.22, X 4 ₌ -0.11, it comes to the X 5 ₌ -0.09 are substituted into equation (8) above and C = 220mg / dl.

Moreover, as an example of the measured value of a healthy person, when the glucose concentration by the enzyme electrode method is 88 mg / dl, the normalization parameters X _N1 = −0.05, X _N2 = + 0.03 obtained from the measurement by this method at the same time, Substituting X _N3 = + 0.06, X _N4 = -0.10, and X _N5 = + 0.12 into the above equation yields C = 94 mg / dl. From the above results, it was confirmed that the glucose concentration can be determined with high accuracy by the method of the present invention.

  FIG. 11 is a graph in which measurement results are plotted for 50 diabetic patients, with the vertical axis representing the glucose function measured with this apparatus by selecting the regression function for diabetic patients and the horizontal axis representing the glucose concentration simultaneously measured by the enzyme electrode method. It is. The correlation coefficient is 0.9473. When a straight line y = Ax + B (where y is a vertical axis and x is a horizontal axis) is fitted to each plot of this figure by the method of least squares, A = 0.992 and B = −6.07. In addition, FIG. 12 plots the measurement results for 50 healthy subjects, with the vertical axis indicating the glucose concentration measured with this apparatus by selecting the regression function for healthy subjects and the horizontal axis indicating the glucose concentration simultaneously measured by the enzyme electrode method. FIG. The correlation coefficient is 0.9388. For each plot in the figure, a straight line y = Cx + D (y is the vertical axis and x is the horizontal axis) is fitted by the least square method, and C = 0.971 and D = 6.84.

  On the other hand, in FIG. 13, the vertical axis is measured with the present apparatus storing the common regression function obtained from the measurement data group of the subject group consisting of the diabetic patient and the healthy person instead of the regression function for the diabetic patient and the regression function for the healthy person. It is the figure which plotted the measurement result about a total of 100 people of 50 diabetic patients and 50 healthy persons as glucose concentration which measured the glucose concentration and the horizontal axis | shaft simultaneously with the enzyme electrode method. The correlation coefficient is 0.9320. For each plot in this figure, when fitting a straight line y = Ex + F (y is a vertical axis and x is a horizontal axis) by the least square method, E = 0.962 and F = 8.15.

  The closer the measured value by this device and the measured value by the enzyme electrode method, the closer the measurement accuracy by this device is to the invasive method. That is, the plots of these values indicate that the closer the correlation coefficient is to 1, the higher the measurement accuracy by this apparatus. Therefore, based on the results of FIGS. 11-13, the common regression obtained from the measurement data group of the subject group consisting of the diabetic patient and the healthy person by appropriately selecting the regression function for the diabetic patient and the regression function for the healthy person and performing the measurement. It can be seen that more accurate measurement is possible compared to measurement using a function.

The model figure explaining the heat transfer from a body surface to a block. Shows the chronological variation of the measured values of the temperature T ₁ and temperature T _2. Measurement example of time variation of the temperature T _3. Explanatory drawing which illustrated the relationship between the measured value by various sensors, and the parameter derived | led-out from it. The top view of the non-invasive blood glucose level measuring apparatus by this invention. Functional block diagram of the apparatus. The figure which shows the detailed example of a measurement part. The conceptual diagram which shows the flow of the data processing in an apparatus. The figure which shows the operation procedure of an apparatus. The figure which shows the detailed operation | movement flow of a selection part. The plot figure of the value calculated by selecting the regression function for diabetic patients and the measured value by the enzyme electrode method. The plot figure of the value calculated by selecting the regression function for healthy persons, and the measured value by the enzyme electrode method. The plot figure of the value calculated by selecting the regression function common to a healthy person and a diabetic patient, and the measured value by the enzyme electrode method.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 11 ... Operation part, 12 ... Measurement part, 13 ... Display part, 15 ... Finger placing part, 16 ... Open end of radiation temperature sensor part, 17 ... Contact temperature sensor part, 18 ... Optical sensor part, 21 ... Plate, 22 ... Thermal conduction member, 23 ... thermistor, 24 ... thermistor, 25 ... infrared lens, 26 ... infrared transmission window, 27 ... thermopile, 28 ... thermistor, 31, 32 ... optical fiber, 33, 34 ... light source, 35 ... photodiode

Claims (11)

A measurement unit for acquiring a plurality of measurement values with respect to the body surface and the measurement environment, and a selection means for selecting a healthy person or a diabetic patient,
A blood glucose level measuring apparatus comprising: a plurality of measurement values acquired by the measurement unit and a calculation unit that calculates a blood glucose level using a selection result by the selection means.   2. The blood sugar level measuring apparatus according to claim 1, wherein the selection means includes a display unit for instructing selection of a healthy person or a diabetic patient.   2. The blood sugar level measuring apparatus according to claim 1, wherein the selecting means includes an input operation unit provided corresponding to each of the healthy person and the diabetic patient.   2. The blood glucose level measuring apparatus according to claim 1, further comprising a storage unit storing a plurality of regression functions, wherein the calculation unit reads the regression function corresponding to the selection result from the storage unit and calculates a blood glucose level. A blood glucose level measuring apparatus characterized by that.   The blood glucose level measurement apparatus according to claim 1, further comprising a storage unit that stores a plurality of regression functions and an average value and a standard deviation of a plurality of parameters corresponding to each regression function, wherein the calculation unit includes the selection result. A blood glucose level measuring apparatus that reads the corresponding regression function, the average value of the parameters, and the standard deviation from the storage unit and calculates a blood glucose level. Input means for inputting the type of healthy person or diabetic patient,
A calorific value measuring unit that measures a plurality of temperatures derived from the body surface and obtains information used for calculating the convective heat transfer amount and the radiant heat transfer amount related to heat dissipation from the body surface;
An oxygen measurement unit that obtains information on the amount of oxygen in the blood;
A function for associating a blood glucose level with a parameter corresponding to each of the plurality of temperatures and the amount of oxygen in the blood, and a storage unit that separately stores a function for a healthy person and a function for a diabetic patient;
The healthy person who has converted a plurality of measured values input from the calorie measuring unit and the oxygen amount measuring unit into the parameters, and stored the parameters in the storage unit according to the type input by the input means A calculation unit for calculating a blood glucose level by applying to a function for the above or a function for the diabetic patient;
A blood glucose level measuring apparatus comprising: a display unit that displays the blood glucose level calculated by the arithmetic unit.   7. The blood glucose level measuring apparatus according to claim 6, wherein the storage unit stores a regression function for healthy persons and a regression function for diabetic patients, and the calculation unit uses a regression function corresponding to the input type. A blood glucose level measuring device characterized by calculating a blood glucose level.   The blood glucose level measuring apparatus according to claim 6, wherein the storage unit stores a regression function for healthy persons, a regression function for diabetic patients, and an average value and a standard deviation of a plurality of parameters included in each regression function. And the said calculating part calculates a blood glucose level using the average value and standard deviation of the said parameter which accompany the regression function corresponding to the said input type, and the said regression function, The blood glucose level measuring apparatus characterized by the above-mentioned.   7. The blood sugar level measuring apparatus according to claim 6, wherein the oxygen amount measuring unit includes a blood flow measuring unit that obtains information relating to a blood flow, and an optical measuring unit that obtains hemoglobin concentration and hemoglobin oxygen saturation in blood. A blood glucose level measuring device. Obtaining a plurality of measurements on the body surface and measurement environment;
Obtaining the type of healthy or diabetic,
A blood glucose level measuring method, comprising: calculating a blood glucose level using a plurality of acquired measurement values and a regression function for a healthy person or a regression function for a diabetic patient. The blood sugar level measuring method according to claim 10, wherein the step of calculating the blood sugar level comprises:
Obtaining a plurality of parameters from a plurality of acquired measurement values;
Normalizing a plurality of obtained parameters with an average value and a standard deviation corresponding to whether healthy or diabetic;
And applying a plurality of normalized parameters to a regression function corresponding to whether the subject is a healthy person or a diabetic patient, and calculating the blood glucose level.

Images